PNP-less Expo current source (NOT sink)

Haible Juergen Juergen.Haible at
Fri Nov 14 12:59:37 CET 1997

Glad that the DIY list seems to work again.
Here's some input:

I looked a little closer at the VCF of the Transcendent Synth (from DIY
last night, and I was surprised to find a very interesting current
source to
control the otas of the filter.

Background: You need a pair of fairly good npn transistors to build a
exponential current *sink* (mostly used in VCOs).

If you want to build Filters (or triangle VCOs) with standard OTAs, an
current *source* would be pleasant. This is where the problem starts, as
transistors are hard to thermo-couple, and *good* integrated pnps are
There's an array (CA3096 ?) that has 3 npn's and 2 pnp's, but I think
the pnp's
are much inferior in quality. (beta, and so on).
I think it was Grant Richter (?) who used them for exponential current
but simply by building a sink from the npns, and using the pnp's as a
(Was that right?)

Here, in the Transcendent, is another way to do it, and it looks very
Sure, it still uses a pnp transistor, but its temperature effects are
because its in an opamp feedback loop, so you can use a discrete one.
You only need npn's for the expo conversion. With an npn array, and the 
multiple opamp/discrete-pnp circuits, you could build *tracking*
current sources. (Important if you want to build a 2pole VCF from
or a 4pole VCF from dual devices like 13700's or 3280's.)

I was surprised to see this elegant circuit, dating back so long as to
the origins
of the Transcendent. (Wasn't this Tim Orr ??)
I don't remember anybody mentioned it on DIY yet.

Does the circuit have any drawbacks ?
Anybody built it ?


More information about the Synth-diy mailing list